It is needed for avoiding nested SIGSEGV on Linux.
Note: BUFSIZ is not proper buffer size. It's unrelated with maximum
filename length. :-/
[Bug #5139] [ruby-dev:44315]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32844 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
gemspecs on install rather than generate fake ones for all gems.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32838 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
C compiler specified by CC variable (e.g. use g++-4.2 for gcc-4.2).
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32836 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This is made the comparison between current and initial commit.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32830 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
'BindAddress: var' => ':BindAddress => var' according to other configurations.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32828 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
[Bug #5153] [ruby-core:38736]
* io.c (argf_next_argv): remove the call of above function.
* ext/-test-/add_suffix, test/-ext-/test_add_suffix.rb: remove the test
extension module because this is only for testsing ruby_add_suffix().
* LEGAL: remove the mention about a part of util.c, because now we
removed the part.
* io.c (argf_next_argv): now the new filename is not guranteed to
use, so should check the return value of rename(2).
* test/ruby/test_argf.rb (TestArgf#test_inplace_rename_impossible):
now we expect same result with other platforms on no_safe_rename
platforms (=Windows).
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32826 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This reverts commit r32817.
It has less code on the surface, but it needs more condition.
Moreover it doesn't need !.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32822 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
[ruby-dev:44315] [Bug #5139]
* vm.c (thread_free): use free because objspace is not ready.
* vm.c (th_init): use malloc because objspace is not ready.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32815 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
use str_to_encindex.
* encoding.c (str_to_encindex): split from to_encoding.
* encoding.c (rb_to_encoding): use str_to_encoding.
* encoding.c (rb_obj_encoding): don't bypass rb_encoding*.
If it uses rb_encoding*, it bypass encindex. If it uses encindex,
it doesn't bypass.
* encoding.c (enc_find): add shortcut for encoding object, use
str_to_encindex, and avoid bypass rb_encoding*.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32811 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
If you connect to localhost, you should listen localhost.
* test/net/http/test_https.rb (TestNetHTTPS#test_timeout_during_SSL_handshake):
ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32806 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
XMLRPC::Client.new3(), when called without host: argument, tries
to connect to a host where "localhost" resolves to. On the
other hand a WEBrick::HTTPServer.new(), when called without
BindAddress: argument, tries to listen all the address where
getaddrinfo(AF_UNSPEC) resolves to. This is a mismatch because
"localhost" might not resolve to one of those listening sockets.
We would better explicitly specify "localhost" here and if
failed, just skip the whole test.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32795 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
--
* enc/Makefile.in (ECHO1): Same as the recent fix in common.mk.
":" in a make variable replacement cause a syntax error with
/usr/ccs/bin/make on Solaris. Uses $(NULLCMD) instead.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32787 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
--
* complex.c (f_signbit): gcc4 on Solaris DOES have signbit but does
not have it on header.
* math.c: ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32786 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
--
* lib/mkmf.rb (configuration:ECHO1): Same as the recent fix in
common.mk.
":" in a make variable replacement cause a syntax error with
/usr/ccs/bin/make on Solaris. Uses $(NULLCMD) instead.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32784 b2dd03c8-39d4-4d8f-98ff-823fe69b080e